To perceive speech is to extract a message. This process may be thought of as a series of hierarchical stages of processing between the initial acoustic waveform and its final conceptual representation. This research proposal is concerned with the earliest stages of linguistic processing--the auditory and phonetic stages. The main goal of this research is the development of a comprehensive theory of speech sound perception. We hope to define the stages of perceptual analysis and specify the types of operations that take place at each of these stages. Central to this work is the use of various "converging operations" within the general framework of human information processing. Techniques and methods from acoustic phonetics and speech synthesis will be used in experiments that range from detection and recognition, dichotic listening, selective adaptation, reaction time, and masking experiments to the study of possible capacity limitations and individual differences. From this work we hope to learn in what ways speech perception may be a "special" process, requiring specialized perceptual and neurological mechanisms and in what ways speech perception may conform to more general perceptual processes.